期刊名称：International Journal of Applied Mathematics and Computer Science
电子版ISSN：2083-8492
出版年度：2010
卷号：20
期号：3
DOI：10.2478/v10006-010-0041-3
出版社：De Gruyter Open
摘要：In this paper we present a method for evaluating the importance of GO terms which compose multi-attribute rules. The rules are generated for the purpose of biological interpretation of gene groups. Each multi-attribute rule is a combination of GO terms and, based on relationships among them, one can obtain a functional description of gene groups. We present a method which allows evaluating the influence of a given GO term on the quality of a rule and the quality of a whole set of rules. For each GO term, we compute how big its influence on the quality of generated set of rules and therefore the quality of the obtained description is. Based on the computed quality of GO terms, we propose a new algorithm of rule induction in order to obtain a more synthetic and more accurate description of gene groups than the description obtained by initially determined rules. The obtained GO terms ranking and newly obtained rules provide additional information about the biological function of genes that compose the analyzed group of genes
关键词：decision rules; importance of rules premises; measures of rules interestingness; gene ontology; descriptions of gene groups
